One aspect of the problem is deeper than just Word, or English. It affects recursively linguistic minorities, and it’s a general tendency with software developers (specially proprietary) and web designers (specially for commercial sites).

I think that it would get 1000% better if those devs and designers followed a few rules, that boil down to “don’t assume”:

1. Don’t assume that languages coincide with countries.
2. Don’t assume that someone from a certain country will want to use your product in a certain language.
3. Don’t assume that your site/software has support for the language that your user would like the most.
4. Don’t assume that the user is or isn’t monolingual.
5. Don’t assume that non-monolinguals are necessarily bilingual, and with English.
6. Don’t assume that the user is wrong or right.

Spelling corrector is a mixed blessing. It has some value as long as the user is aware of its shortcomings, and it should always worded in a way that acknowledges those shortcomings. e.g.

• [Bad example] «wrod» is likely wrong. Do you mean «word»?
• [Better example] «wrod» was not recognised. Do you mean «word»?
• [Even better] «wrod» was not recognised. Similar word: «word»

I think the examples in the article are a bit too high level, although accurate - even more interesting when they affect grammar, like both MS Office and Grammarly leading a crusade against the passive voice.

More interesting to me though is how Microsoft Windows (not just Office) lead to the extinction of a whole punctuation point in my native Greek. The “Greek semicolon” was not included in the default Greek keyboard layout for Windows. While it remained as an option on the IBM keyboard that big organisations could choose to order, it vanished from retail and therefore from home users and the language simply lost an entire punctuation mark within a decade.

If there’s a clear example of how technology can drive language change (to the extend that writing is part of language), I feel like that’s one of the clearest examples.